The main goal of this research program is to characterize the development of the mouse brain, with emphasis on white matter anatomy, using magnetic resonance micro-imaging in conjunction with mathematical methodologies for quantitative image and shape analysis following the field of computational neuroanatomy. The development of methods for altering the genotype of model animals, including mice, has opened enormous possibilities for better understanding the role of different genes in brain development and diseases, by studying differences in the course of brain development between normal and transgenic animals. Imaging, and in particular MRI, is bound to play an important role in structural phenotyping of mouse models, since it can serve as an effective screening tool pointing to more detailed yet laborious histological analyses, and does not suffer from sectioning and staining distortion artifacts, but provides spatially consistent 3D volumes. However, conventional imaging techniques cannot reveal the internal white matter architecture, which consists of various fiber tracts and is of great interest in understanding brain development. This is especially the case for fetal and young mouse brains, in which myelination is not well developed and the T1- or T2-weighted images cannot even distinguish the gray and white matter with adequate contrast. Diffusion tensor imaging (DTI) provides excellent contrast, and is able to distinguish various structures in the white matter of even young brains, by imaging microscopic water diffusion which is known to be relatively higher along axonal fibers. Therefore, DTI is bound to be critical in studying brain development and brain connectivity, as well in understanding how genetic mutations affect the formation, myelination, or degeneration of axonal fibers. Concurrent with advances in micro-MR imaging have been advances in mathematical methodologies for computational anatomy, a rapidly maturing field of quantitative techniques for analysis of brain morphology from volumetric images. These new approaches are now adopted by a number of neuroimaging and neuroinformatics groups, and they constitute powerful tools for structural phenotyping in highly automated and detailed ways; they are particularly sensitive to subtle and spatially complex patterns of morphological change. In this project, we will continue to develop methods for quantitative analysis of the mouse brain, and use them to generate normative data for brain development of the C57BL/6J mouse strain, and to investigate phenotypic differences between wildtype and transgenic mouse models. Emphasis will continue to be the mathematical and computational challenges posed by the complexity of tensor images, which are obtained in DTI (Aims 1 and 2), especially challenges in image registration and morphological analysis that are posed by the rapid changes observed during early postnatal development as well as by morphological differences between transgenic and wildtype mice. Moreover, we will test the utility of our methodologies in studies of 3 specific projects involving mouse models of schizophrenia, autism, and Rett's syndrome (Aim 3), in collaborative efforts.